* Reapply "More aggressive using of cache on engine settings changes (#4691)" (#4729)
This reverts commit 3f1f40eeba.
* Add a utility to get all the current values from the settings object
* Use an XState selector to get the latest settings snapshot for WASM
---------
Co-authored-by: Frank Noirot <frank@kittycad.io>
* Parse more import syntax
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* Remove unnecessary Vec from VariableDeclaration
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* Parse export import
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* Factor out an execution module
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* imports: constants, globs, export import
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* test fixups
Signed-off-by: Nick Cameron <nrc@ncameron.org>
---------
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* Fix argument error to point to the arg at the call site
* Change MemoryFunction to accept Vec<Arg> instead of Vec<KclValue>
* Rename variable to be clearer
* Fix one more argument error message
* Clean up from PR feedback
* Make ast module private
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* Make most other modules private
Signed-off-by: Nick Cameron <nrc@ncameron.org>
* Expand API to support CLI, Python bindings, and LSP crate
Signed-off-by: Nick Cameron <nrc@ncameron.org>
---------
Signed-off-by: Nick Cameron <nrc@ncameron.org>
Removes JSON from the KCL object model. Closes https://github.com/KittyCAD/modeling-app/issues/1130 -- it was filed on Nov 27 last year. Hopefully I close it before its one year anniversary.
Changes:
- Removed the UserVal variant from `enum KclValue`. That variant held JSON data.
- Replaced it with several new variants like Number, String, Array (of KCL values), Object (where keys are String and values are KCL values)
- Added a dedicated Sketch variant to KclValue. We used to have a variant like this, but I removed it as an experimental approach to fix this issue. Eventually I decided to undo it and use the approach of this PR instead.
- Removed the `impl_from_arg_via_json` macro, which implemented conversion from KclValue to Rust types by matching the KclValue to its UserVal variant, grabbing the JSON, then deserializing that into the desired Rust type.
- Instead, replaced it with manual conversion from KclValue to Rust types, using some convenience macros like `field!`
* Add ts_rs feature to work with indexmap
* Add feature for schemars to work with indexmap
* Add module ID to intern module paths
* Update code to use new source range with three fields
* Update generated files
* Update docs
* Fix wasm
* Fix TS code to use new SourceRange
* Fix TS tests to use new SourceRange and moduleId
* Fix formatting
* Fix to filter errors and source ranges to only show the top-level module
* Fix to reuse module IDs
* Fix to disallow empty path for import
* Revert unneeded Self change
* Rename field to be clearer
* Fix parser tests
* Update snapshots
* Change to not serialize module_id of 0
* Update snapshots after adding default module_id
* Move module_id functions to separate module
* Fix tests for console errors
* Proposal: module ID = 0 gets skipped when serializing tokens too (#4422)
Just like in AST nodes.
Also I think "is_top_level" communicates intention better than is_default
---------
Co-authored-by: Adam Chalmers <adam.chalmers@zoo.dev>
Demonstrate simulation tests where we don't care about visuals, e.g. the double-map test.
The `just new-sim-test` now accepts an optional argument, `render_to_png` which can be either "true" or "false" (defaults to "true"). Tests like double_map that don't render anything can use false, rather than rendering an empty PNG with nothing in it.
This means the [tests under `no_visuals/`](https://github.com/KittyCAD/modeling-app/tree/v0.26.2/src/wasm-lib/tests/executor/inputs/no_visuals) can be entirely replaced by simulation tests. This is much better! For example, I moved `double_map.kcl` from a no_visuals test to a simulation test. Here's the file:
```
fn increment = (i) => {
return i + 1
}
xs = [0..2]
ys = xs
|> map(%, increment)
|> map(%, increment)
```
Previously the `no_visuals` test just checked that the program ran successfully without panicking. Now the simulation test lets you see the value of `xs` and `ys` and immediately see they're correct. If our map logic changes (for example, we have an off-by-one error and don't apply the `map` to the last element) it'll show up in the program memory snapshot.
* Add lsystem.kcl to tests
* Reduce iterations
* Fix the user settings flake shit (NOTE TO ALL FUTURE PEOPLE MODELING-APP DOES NOT WAIT FOR I/O IN SOME CASES BEFORE ROUTER NAVIGATION)
Josh Gomez requests pattern calculations take the total number of instances,
not the number of extra repetitions to do. This is how we designed the
patternTransform API, but we didn't do that for patternLinear/Circular.
* Log any Errors to stderr
This isn't perfect -- in fact, this is maybe not even very good at all,
but it's better than what we have today.
Currently, when we get an Erorr back from the WebSocket, we drop it in
kcl-lib. The web-app logs these to the console (I can't find my commit
doing that off the top of my head, but I remember doing it) -- so this
is some degree of partity.
This won't be very useful at all for wasm usage, but it will fix issues
with the zoo cli silently breaking with a "WebSocket Closed" error --
which is the same issue I was solving for in the desktop app too.
In the future perhaps this can be a real Error? I'm not totally sure
yet, since we can't align to the request-id, so we can't really tie it
to a specific call (yet).
* add to responses
Signed-off-by: Jess Frazelle <github@jessfraz.com>
* A snapshot a day keeps the bugs away! 📷🐛 (OS: ubuntu-latest)
* add a test
Signed-off-by: Jess Frazelle <github@jessfraz.com>
* clippy[
Signed-off-by: Jess Frazelle <github@jessfraz.com>
* A snapshot a day keeps the bugs away! 📷🐛 (OS: ubuntu-latest)
* empty
* fix error
Signed-off-by: Jess Frazelle <github@jessfraz.com>
* updates tests
Signed-off-by: Jess Frazelle <github@jessfraz.com>
* docs
Signed-off-by: Jess Frazelle <github@jessfraz.com>
---------
Signed-off-by: Jess Frazelle <github@jessfraz.com>
Co-authored-by: Jess Frazelle <github@jessfraz.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Jess Frazelle <jessfraz@users.noreply.github.com>
Previously variable declaration required a keyword, e.g.
```kcl
let x = 4
const x = 4
var x = 4
```
These were all valid, and did the exact same thing. As of this PR, they're all still valid, but the KCL formatter will change them all to just:
```kcl
x = 4
```
which is the new preferred way to declare a constant.
But the formatter will remove the var/let/const keywords.
Closes https://github.com/KittyCAD/modeling-app/issues/3985
* circle
* fix another example
* fix bad comment
* toPoint fix
* cargo fmt
* resolve most of the tests
* fix last test
* missed circle in bracket
* remove console error
* fmt
* A snapshot a day keeps the bugs away! 📷🐛 (OS: ubuntu-latest)
* A snapshot a day keeps the bugs away! 📷🐛 (OS: windows-latest)
* trigger ci
* remove three dot menu for circle
* make sure circle can be extruded
* fix up after merge
* add extrude test for circle
* clean up
* typo
* A snapshot a day keeps the bugs away! 📷🐛 (OS: windows-latest)
* Revert "A snapshot a day keeps the bugs away! 📷🐛 (OS: windows-latest)"
This reverts commit 03f8eeb542.
* update docs again
* cmd bar test serialisation improvements
* tiny clean up
* fix after: Replace kittycad crate with kittycad-modeling-cmds
* fmt
* rename fix
* Update src/lib/toolbar.ts
Co-authored-by: Frank Noirot <frank@zoo.dev>
* add another error to list
* A snapshot a day keeps the bugs away! 📷🐛 (OS: ubuntu-latest)
* image updates
* Revert "A snapshot a day keeps the bugs away! 📷🐛 (OS: ubuntu-latest)"
This reverts commit 505bb20bea.
* update markdown
* skip un reproducable windows test failure
* rust review
* leave issue todo comment
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Frank Noirot <frank@zoo.dev>
Now the process for adding a new KCL test is:
- `cd src/wasm-lib`
- Write a new .kcl file in the `tests/inputs` directory, e.g. `tests/inputs/foo.kcl`
- Run `just new-test foo`
- That created `tests/outputs/foo.png`, so open it and check it looks correct
- Commit changes and open a PR
* updated tangentArc math
* Add a test case showing tan arc then xLineTo
* Fix compile errors
* Tweaking the math
* Use + on angles
* atan2 outputs radians, not degrees
* Track ccw and center of all tan arcs
* re-sequenced atan2 arcTan
* Remove print statements
* Update the test
* Update kittycad in tauri
* New arc fields
---------
Co-authored-by: Jordan Noone <jordan@kittycad.io>
This new test framework, `no_visuals`, is for testing KCL programs via asserts, not via twenty-twenty visual tests. This is useful for unit-testing small fragments of KCL.
It's easy! All you need to do is:
- Write a KCL file
- Save it under `tests/executor/inputs/no_visuals/foo.kcl`
- Open `no_visuals.rs` and add `gen_test!(foo);`
'serial_test' isn't actually accurate. Two of these tests run in parallel
now. So I renamed it 'kcl_test' as that's what it's actually doing.
In the nextest config, I changed the label from 'serial-integration' to
'uses-engine' because the former isn't true, and also doesn't explain
_why_ it's being limited. The new name explains why we're limiting the
number of tests that can run in parallel.
* Use kcl_input! macro
These two lines are equivalent:
`kcl_input!("foo")`
`include_str!("inputs/foo.kcl")`
Simplifies the tests.
* Move more KCL test programs into their own files
* Move twenty-twenty asserts into their own function
* Move more asserts into 'assert_out'
* Fix to allow variable shadowing inside functions
* Implement closures
* Fix KCL test code to not reference future tag definition
* Remove tag declarator from function parameters
This is an example where the scoping change revealed a subtle issue
with TagDeclarators. You cannot bind a new tag using a function
parameter.
The issue is that evaluating a TagDeclarator like $foo binds an
identifier to its corresponding TagIdentifier, but returns the
TagDeclarator. If you have a TagDeclarator passed in as a parameter
to a function, you can never get its corresponding TagIdentifier.
This seems like a case where TagDeclarator evaluation needs to be
revisited, especially now that we have scoped tags.
* Fix to query return, functions, and tag declarator AST nodes correctly
